Block 3: Week 1: Epidemiology Flashcards
What is Epidimeology?
Study of disease in populations
Define Prevalance
Number of people with a problem in a defined population at one time
Define Indicidence
Number of new cases of a problem in a defined population in a defined period of time
Define mortality rate
Number of people dying in a defined population in a defined period of time
What is causality?
Differentiating association from causation
Concepts of causality: what are the two types?
Deterministic approach (A causes B)
and
Stochastic approach (likelihood/risk)
Concepts of causality: Deterministic approach
Deterministic inevitability
Validation of hypothesis by systematic observations to predict with certainty future events
Concepts of causality: Stochastic approach
Ax of hypothesis by systematic observations to give risk of future events
Concept of causality: Deterministic approach- what are some of the features? What is it useful for?
Newtonian thinking (A leads to B)
Mechanistic, can take apart to study
Objective, quantifiable and certain
Whole is the sum of the parts
Useful:
Single cause for a single disease
Concept of causality: Stochastic approach- what are the features?
Quantum thinking (Hollistic approach, not just linear)
Whole greater than sum of parts
Whole not predictable from knowledge of parts
Probabilities of certainties
Systems theory; complexity theory
- The observer influences the observed
- Emergent phenomena
Confouding: What are confounding factors?
Something that is associated with both the exposure & the outcome
Eg: Smoking is associated with high blood pressure (exposure) and MI (outcome)
Confouding: What is exposure?
An exposure is independently associated with the outcomes even after taking confounding factors into account
(Eg: Smoking causes MI by itself. It can also cause increased BP which can calso cause MI.
Smoking = confouding factor
BP = Exposure)
Confouding: What is a mediating variable?
A variable through which an exposure wholly or partially exerts its effect
eg: Eating sugar (exposure) causes obesity (mediator) which causes heart disease
Confouding: What is reverse causality?
Both factors can cause each other
Eg: Loosing a job causes mental illness
BUT
People with mental illness can cause unemployment
What is Causality: Aetiology?
Epidemiology is very good at Ax the disease risk associated with individual agents
In assess probability, epidemiological studies CANNOT prove causality but can make a case ‘beyond reasonable doubt’
What is Bradford Hill’s Criteria for Inferring Causality: Assoication factors? (1965)
1) Strength of association: A causal link is more likely with strong associations (measured by rate ratio OR odds ratio) & are unlikely to be explained by bias or confounding factors.
Eg: heavy smokers x20 likely of mortality from laryngeal ca tha non-smokers. Weak association can be causal.
2) Specificity of association: A causal link is more likely when a disease is assoicated with a specific factor & vice-versa. Eg: asbestos –> mesothelioma. Lack of specificity does not necessarily weaken the case. Current models of disease are multi-factoral.
3) Consistency of association: A causal link is more likely if the assoication is observed in different studies & subgroups.
- Less likely to be due to bias/ confounding factor
- Lack of consistency can be due to study design
Bradford Hill’s Criteria for Inferring Causality (1965): What are the exposure and outcome features?
1) Temporal Sequence: A causal link is more likely if the exposure to the putative cause has been shown to precede the outcome.
- Converse: causal link cannot exsist if the outcome precedes the esposure to putative factor
- Optimal study deigns: Prospective cohort/ RCT
- Weak study designs: Cross-sectional/ Case control study
2) Dose response (biological gradient): A causal link is more likely if different levels of exposure to the putative factor lead to different risk of acquiring the outcome. Eg: 5 a day.
* Unlikely to be due to unknown confounding or bias
3) Reversibility: A causal link is very likely if removal or prevention of the putative factor leads to a reduced or non-existent risk of acquiring the out comes.
- Probably strongest evidence for causal link but dificult to demonstrate
- Disease: long time lags
- Ethical issues for RTC
- Public health programme require action by society
Bradford Hill’s: What are the three other evidences?
1) Coherence of Theory: A causal link more likely if the observed association conforms with current knowledge- paradigms/ constructs/ theories
- Lack does not rule out causal link
- Inapproriate rejection of ‘unfavoured’ associations
8) Biological Plausibility: A causal link is more likely if a biological plausibility mechanism is likely or demonstrated
* eg: babies sleeping on tummy
9) Analogy: If an analogy exists with other disease, species or settings
* Easier to infer than a biologically plausible mechanism
Epidemiological study designs: What are the observational studies, what are they used for?
1) Cross sectional Survey: May be set up for specific purpose
- Prevalance of specific disease
- Investigate distribution of a specific disease in population
- Monitoring health over time
- Medium cost
2) Case control Studies: Almost always set up for speicifc purpose - Investigate suspected determinants eg: Outbreak investigation/ determinants of rare conditions
- Quick & Cheap
- Recall and selection bias an issue
3) Cohort: May be for specific purpose. Often multipurpose
- Eg: Effects of smoking, asbestos
- Determinants of common conditions
- Relative importance of different factors
- (Very) Slow & (very) Expensive
- Issues due to confounding with unknown risk factors
Epidemiological study design: Experimental studies- what are they and what are they used for?
1) Uncontrolled Studies
- Easy to do
- Selection bias a problem
- Don’t know what happens without intervention
2) Natural Experiments:
* May be only way to investigate some things due to unkown confounding factors
3) Controlled studies
4) RTC
What is the hierarchy of evidence?
Systematic reviews
Experimental studies
- Randomised Controlled Trials
- Controlled trials
Observational studies
- Cohort studies
- Case control studies
Descriptive studies
- Cross sectional
- (Qualitative studies)
What is bias?
Any trend in the collection, analysis, interpretation, publication of review of data than can lead to conclusions that are systematically different from the truth
What are the types of bias in epidemoilogical studies?
1) Selection (design plus execution)
Admission, prevalence/ incidence, detection, volunteer, loss to follow up
2) Information (data collection)
Interviewer, questionaire, recall, diagnostic suspicion & recall
3) Confounding
What are some of the epidemiological issues?
They provide information of the average effect
Average effect hides individual level
For some patients better not to do what is best of average- patient preference support these
When are epidemiological studies best, good and more limited?
Best:
- Single agent causes disease
- Single Tx reverses disease
Good:
- Primary factor causes disease with several secondary influences
Limited:
- Many different factors interact in complex pathways to create the conditions in which multiple diseases are likey to arise eg: social inequalities & health
